Lessons Learned - great idea, but its in the wrong place!

Why do we teach project managers to include the Lessons Learned (LL) review in the Close Down phase of their project?
and
Why do PM's then seldom conduct LL reviews?

Simply; its in the wrong place.

Yes, semantically we can only learn the lessons of history once they have happened, but I argue we should be teaching PM's to include the Lessons Learned review in the Initiation phase of their project, AT THE START.

The reasons for this approach are two-fold:
(1) The raison d'etre of LL reviews is that the lessons can be applied to future projects, but do we ever formally capture this in our project methodologies? Yeah its there somewhere, but is often only given lip-service. But ask any PM "When do you conduct your LL review on a project?" and they'll fire back "during the Close Down phase!". The emphasis is in the wrong place. So simply put, if we put the LL review in the initiation phase of the project, the whole point of doing them will be realised. OK, so the format probably needs to change and we need to consider that project teams quickly disband after a project finishes, so we need to capture their experiences when they are fresh. However in todays Web 2.0 environment, call-it-what-you-will, the ability to contact people and extract information from them has never been easier. However, in fairness I still see the LL review being conducted at the end of the project too, but I foresee that they will be conducted more readily, if we change the emphasis of when we do them.
(2) To enlarge on this point; if the PM's actively conduct LL reviews at the start of their projects, they will vividly see the benefits of this activity. Ergo, they will also see the benefit in providing the Lessons Learned after their own project finishes and the whole process will become more self-sustaining. Let's face it, any who has ever conducted a Delivery/Quality Assurance review, or sat on a Governance Board, will know that happy projects that have run successfully and dire failures are they only two where LL reviews are typically mandated. All the projects in the middle ground that struggle and take up most of people's bandwidth, often get closed in a hurry as people run for the door into another similar project to repeat the cycle.

Challenge everything, accept nothing, improve something.